Transfer Learning across Several Centuries: Machine and Historian Integrated Method to Decipher Royal Secretary's Diary

TL;DR

This study develops a specialized Korean historical corpus with annotated entities, fine-tunes language models for NER, and demonstrates that phrase markers significantly enhance the detection of unseen entities in centuries-old documents.

Contribution

Introduces a Korean historical corpus with annotated entities, and shows that phrase markers improve NER performance on historical texts.

Findings

Phrase markers improve NER accuracy on unseen entities.

Corpus-specific models alone do not outperform pretrained models.

Combining time and annotation info benefits historical text analysis.

Abstract

A named entity recognition and classification plays the first and foremost important role in capturing semantics in data and anchoring in translation as well as downstream study for history. However, NER in historical text has faced challenges such as scarcity of annotated corpus, multilanguage variety, various noise, and different convention far different from the contemporary language model. This paper introduces Korean historical corpus (Diary of Royal secretary which is named SeungJeongWon) recorded over several centuries and recently added with named entity information as well as phrase markers which historians carefully annotated. We fined-tuned the language model on history corpus, conducted extensive comparative experiments using our language model and pretrained muti-language models. We set up the hypothesis of combination of time and annotation information and tested it based on statistical t test. Our finding shows that phrase markers clearly improve the performance of NER model in predicting unseen entity in documents written far different time period. It also shows that each of phrase marker and corpus-specific trained model does not improve the performance. We discuss the future research directions and practical strategies to decipher the history document.